Conceptual design of hybrid battery modules for optimising electric aircraft performance

(1)

C ONCEPTUAL D ESIGN OF H YBRID

B ATTERY M ODULES FOR O PTIMISING

E LECTRIC A IRCRAFT P ERFORMANCE

Gabriel-André Dominic Damian Bachelor Thesis

Chair: Prof.dr.ir. Braham Ferreira

Supervisor: M. Sc. Alexander Matthee, PhD Student

Submitted in fulfilment of the requirements for the degree of Electrical Engineering Bachelor

Power Electronics & Electromagnetic Compatibility (PE)

Electrical Engineering, Mathematics and Computer Science Department (EEMCS) University of Twente

Academic Year 2019/2020

(2)

(3)

Keywords

A3T, Aerobotic Tech Team Twente, Alice, Battery, EA, Electric Aicraft, Ehang184,

Energy Density, EV, Electric Vehicle, Eviation, Fixed Wing, Full Electric, Hybrid

Battery, Lithium polymer, Lithium-ion, Load Switch Circuit, Multicopter, Power

Density, Specific Energy, Specific Power, UAM, VAU, VTOL,

(4)

ii Conceptual Design of Hybrid Battery Modules for Optimising Electric Aircraft Performance

Abstract

The presented research investigates the usage of a hybrid battery solution for full-electric aircraft as mean of battery weight optimization. The concept is to utilize a high-energy dense battery module for the cruise phase and a high-power dense battery module for the take-off and climb phase, rather than to use a singular battery for the whole flight. It was previously expected that such approach would overall increase the specific energy of the battery without reducing the power performance for take-off, hence reducing the weight of the battery.

The result, however, pointed out that a hybrid solution will in most cases lead to a minor weight increase of the battery, but will also significantly increase the energy density. It was found that a hybrid solution inherently cost additional weight for an- extra power-dense battery module for the take-off and climb phase.

Overall, a hybrid solution is not recommended for long range fixed-wing configuration. But for multicopter and VTOL, a hybrid solution is recommended as it may increase the operational range by 63% and 84% with only 7% and 16% battery weight increase respectively.

Finally, the research has found that the optimal circuit topology for switching

between two battery sources is best achieved using N-Chanel power MOSFETs in

parallel and in dual configuration. With this approach, a scalable circuit can be

developed to handle high-power requirements with minimal energy losses.

(5)

Chapter 1: Introduction ... 1

1.1 Background ... 1

1.2 Context ... 2

1.3 Purposes ... 2

1.4 Significance, Scope and Definitions ... 2

1.5 Thesis Outline ... 4

Chapter 2: Technology Landscape Review ... 5

2.1 Electric Vehicles ... 5

2.2 Energy Storage and batteries ... 8

2.3 Load Switching Circuit... 11

2.4 Summary and Implications ... 12

Chapter 3: Design and Simulation ... 13

3.1 Methodology and research design ... 13

3.2 Software and Instruments ... 16

3.3 Procedure and Timeline ... 17

3.4 Analysis ... 18

3.5 The application... 18

Chapter 4: Results & Validation ... 25

4.1 Alice ... 25

4.2 Ehang 184 ... 26

4.3 VAU ... 28

Chapter 5: Parametric Analysis ... 29

5.1 The results... 29

5.2 Aicraft Performance review ... 31

5.3 Load Switching Circuit... 32

Chapter 6: Conclusions and Future Work ... 33

(6)

iv Conceptual Design of Hybrid Battery Modules for Optimising Electric Aircraft Performance

Bibliography ... 35

Appendices ... 39

Appendix A ... 39

Appendix B ... 40

Appendix C ... 43

Appendix D ... 44

Appendix E ... 45

(7)

List of Figures

Figure 2-1: Hybrid battery powertrain [30] [31] ... 8

Figure 2-2: Ragone chart for various energy storage types [37] ... 10

Figure2-3: Switch on the left, Relay on the right (symbols taken from Sparkfun) ... 11

Figure 2-4: From left to right: Thyristors, BJT, JFET, MOSFET, IGBT symbol (symbols taken from Sparkfun and EasyEDA) ... 11

Figure 2-5: Dual MOSFETs configuration to prevent reverse current [40]... 12

Figure 3-1: Alice aircraft from Eviation [41] ... 14

Figure 3-2: EHANG 184 [26]... 14

Figure 3-3: VAU, designed by members from A3T (courtesy of A3T) ... 15

Figure 3-4: Energy & Power Profile on the left and Battery Profile on the right ... 19

Figure 3-5: Simplified Flight Profile of a Fixed Wing ... 23

Figure: 3-6: Simplified Flight Profile of a Multicopter on the left and VTOL on the right ... 24

Figure 4-1: Range performance relative to battery weight ... 25

Figure 4-2: Results of the parametric analysis for the fix wing configuration ... 25

Figure 4-3: Range performance relative to battery weight for the multicopter ... 26

Figure 4-4: Results of the parametric analysis for the multicopter configuration ... 26

Figure 4-5: Battery weight reduction achieved at different cruise speed for Ehang184 ... 27

Figure 4-6: Battery weight reduction achieved at different range speed for Ehang184 ... 27

Figure 4-7: Results of the parametric analysis for the VTOL configuration ... 28

Figure 4-8: Range performance relative to battery weight for the VTOL

configuration ... 28

(8)

vi Conceptual Design of Hybrid Battery Modules for Optimising Electric Aircraft Performance

List of Tables

[The List of Tables can be created automatically and updated with the F9 key – refer to Thesis PAM.]

Table 2-1: Recent EA designs, [22] [23] [24] [25] [26] [27] [28] [29] ... 7

Table 3-1: Variables studied ... 16

Table 4-1: Weight and capacity comparison ... 25

Table 4-2: Weight and capacity comparison ... 28

(9)

List of Nomenclature

ASSB: All Solid-State Battery CDCL: Climb Drag Coefficient CDCR: Cruise Drag Coefficient CDDC: Descend Drag Coefficient CDTO: Take-Off Drag Coefficient EA: Electric Aircraft

EV: Electric Vehicle

ICE: Internal Combustion Engine TOW: Take-Off Weight

UAM: Urban Air Mobility

VTOL: Vertical Take-Off and Landing

(10)

viii Conceptual Design of Hybrid Battery Modules for Optimising Electric Aircraft Performance

Statement of Original Authorship

The work contained in this thesis has not been previously submitted to meet requirements for an award at this or any other higher education institution. To the best of my knowledge and belief, the thesis contains no material previously published or written by another person except where due reference is made.

Signature: _________________________

Date: _________________________

(11)

Acknowledgements

Over the course of 12 weeks, a considerable amount of time and effort went into the realisation of this research. However, it would not have been possible without the kind (and patient support) of few individuals.

First, I would like to thank Akarsh for sharing his impressive pool of knowledge and understanding in the field of aeronautic. Secondly, for Alex to guide me through this journey and providing me with precious feedback. Thirdly, for Braham for accepting my first research proposal and supporting the pursuit of my passion.

And of course, a special thank you for my mother who supported me

throughout my engineering studies, from mechanical to electrical and from Canada to

the Netherlands.

(12)

(13)

Chapter 1: Introduction

To facilitate the development of full-electric aircraft (EA) for a greener future, the following research investigates how an EA’s battery system can be further optimised using current technology. The focus of the research is to evaluate the conceptual design of a hybrid battery system solution consisting of two modules. The first module with high-energy density, optimised for flight endurance, and the second module with high-power density optimised for power requirements during take-off and climb. If proven successful, such a solution may result in overall lower aircraft battery weight, and thus could increase the EA’s performance. Furthermore, the study will investigate scalable switching circuit approaches that could be used in the realization of such hybrid battery modules to ensure minimal energy loss.

The first section of this chapter covers the basis that led to this research (section 1.1), its context (section 1.2) and its purposes (section 1.3). The significance and scope are then outlined (section 1.4) and the main terms are defined. Finally, a general overview of the following chapters is presented (section 1.5).

1.1 BACKGROUND

“Humanity is now standing at a crossroads. We must now decide which path we want to take” (Greta Thunberg, 2019). As of 2020, it is now well agreed by the scientific community [1] and general public that human is responsible for causing global warming and that climate action is necessary. This realisation is now fuelling the path to the electrification of the future and the energy revolution essential to curb carbon dioxide emissions.

Almost all carbon dioxide emissions originate from the combustion of fossil fuel

[2], therefore it is only logical that reducing the use or finding an alternative to fossil

fuel is an effective way to lower emissions. One way is to electrify means of

transportation, which is responsible for about 20% of global emissions [3], for which

2% originates from the aviation sector [4]. Although the impact of aviation is relatively

small for now, this number is expected to considerably grow as aviation intensifies and

other emission sources such as energy production make rapid progress to decarbonise

[5]. Moreover, one-third of the operating costs of airlines is spent on fuel [6]. With the

(14)

2 Chapter 1: Introduction

continuous rise in fuel prices, the financial incentive is pressuring the industry to focus on fuel efficiency and find alternatives such as full electric flying, as it could cut the direct operating cost since electricity is cheaper than jet fuel.

1.2 CONTEXT

While battery technology for land-based vehicles is now adequate to compete against their ICE counterpart, there is still a long way to go before electric aircraft can do the same. This difficulty mainly arises from the weight attributed to power the vehicles. In comparison, lithium-ion cells approximately hold 2% of the energy density of jet fuel. Even after considering the loss in efficiency due to the thermodynamic cycle, jet fuel yield about 14 times more energy than a lithium-ion battery at equivalent weight [7]. Moreover, unlike cars, weight is a primary concern for performances [8], and thus the electrification of aircraft is not as straight forward as it would be for a car.

Before electric aircraft become viably and financially competitive to conventional aircraft, further development is required to optimise the performance of batteries. For this reason, the presented research will evaluate performance optimisation of batteries based on combining two separate battery modules, each optimised for different phases of flight.

1.3 PURPOSES

The central objective of this thesis is to evaluate if a hybrid battery solution can reduce the overall weight of an aircraft’s battery, and if so, to determine to which extent. The secondary objective is to offer a possible electric circuit design to realize such a concept.

Through the thesis, readers will acquire the basic knowledge and foundation to further research on fully electric hybrid battery solutions.

1.4 SIGNIFICANCE, SCOPE AND DEFINITIONS

Hybrid battery solutions for electric aircraft has been mentioned multiple times

in previous studies but has not been directly studied nor as the practical implications

been evaluated. This research aims to cover this gap in the literature and further expand

the foundation for the development of future full electric aircraft.

(15)

The scope of the research limits itself to the case studies of three aircraft configurations and fundamental electronics. For all models, the analysis will be based on simplified aerodynamics models and conducted through a dedicated MATLAB application. The proposed switching circuit configuration will also be kept simplified by using the minimum of components required to function.

Since definitions of terms may vary from research to research, it is best to clarify the meaning of the main recurring terms used in this study:

• Fixed-wing refers to aircraft using wings and forward propulsion as their only mean of sustaining flight. They are mostly known as airplanes. This type of configuration generally offers the best range and flight endurance.

• Multicopter refers to aircraft using upward propulsions as their only mean of sustaining flight. This type of configuration generally offers the least range and endurance but can manoeuvre in all directions and take off vertically.

• VTOL, or Vertical Take-Off and Landing, refers to aircraft using upward propulsion as mean to take off, and wings with forward propulsion to sustaining flight at cruise. They form a hybrid between fixed-wings and multicopters, resulting in both high manoeuvrability, long-range and endurance.

Moreover, the two main factors used in the study to qualify a battery’s performance are the following:

• Energy density (or specific energy) refers to the energy-to-weight ratio, typically in Watt-hour (Wh) per kilogram (kg). This can be compared to the EA flight range relative to its battery weight.

• Power density (or specific power) refers to the power-to-weight ratio, typically in Watt (W) per kilogram (kg). This describe the available maximum power that the EA can draw per battery unit weight.

For abbreviations, refers to the list of abbreviations.

(16)

4 Chapter 1: Introduction

1.5 THESIS OUTLINE

The presented research follows a standard thesis outline, with a total of six chapters.

• Chapter 1 introduces the motives behind the thesis and its content.

• Chapter 2 reviews the current technology landscape related to the study.

• Chapter 3 outlines the design and methodology of the research.

• Chapter 4 provides the results of the study without interpretations.

• Chapter 5 discusses the parametric impact of variables issued from the results.

• Chapter 6 concludes the research and points to further possible studies.

Further information, additional references, and results can be found in the

appendixes.

(17)

Chapter 2: Technology Landscape Review

There is over a century of research and development that has led to the current state of the art aeronautic technology required to propel modern airplanes in the sky [9]. Since electric aircraft relies on this already well establish bases, the scope of this research will not accommodate for complex aerodynamic studies. Rather, the focus of the research is brought on the emerging electric advancement that has led to the resurgence of electric vehicles.

The first section (2.1) will evaluate the broad nature of electric vehicles, from the ground to the sky. The second section (2.2) will investigate the facets of energy storage, with a closer look to batteries and its limitation. The third section (2.3) will introduce different types of load switching circuit techniques. Finally, the last section (2.4) will summarise the technology landscape review and offer its further implications for the remaining of the study.

2.1 ELECTRIC VEHICLES

Electric vehicles (EV) are vehicles that use electric motors as a mean of propulsion. The electricity required for the motors can come from various sources, such as batteries, electric generator, solar panels and hydrogen fuel cells among others.

Their debuts dates to the mid-19

^th

century, when electricity used to provide superior ease of operation to power vehicles compared to gasoline engines of the time.

In comparison, around one third of all vehicles on the road in the U.S. were electric in the early 20

^th

century. Electric vehicles then saw a decline in 1920 as soon as gasoline became cheaper and readily accessible with the venue of gasoline stations [9].

2.1.1 Land Vehicle

Currently, electric vehicles are gaining in popularity as the focus on renewable energy has pushed new advancement in battery technology and economics are more favourable. However, it wasn’t always the case.

Due to the traditionally high cost of batteries, full-electric cars were relatively

expensive and were out of reach for most of the population. Though, the concept of

electric cars became more known in the late 1990s with the first mass-produced hybrid

(18)

6 Chapter 2: Technology Landscape Review

vehicle, the Toyota Prius [10]. Back then, hybrid vehicles worked primarily on gas engines and only used autogenerating to recharge a small battery pack. The pack was then used for quick acceleration of the car to considerably reduce the fuel consumption in urban stop-start conditions. Regenerative breaking also appeared with the mass production of the Prius. It was only in 2008 that the regenerative braking technology was used in other hybrids as Ford licenced the technology from Toyota [11]. Finally, it’s only later that hybrid used a larger battery pack that could be recharged at home [9].

Since batteries do not have an unlimited serviceable life, various researches have come up with solutions to prolong the EV’s battery life. For example, it is recommended to not charge the car to full capacity, nor to fully discharge it [12] and to slowly charge and discharge the battery to prolong its life [13] .

However, these solutions may not always be adequate for EA. Not using the full capacity of the aircraft’s battery is non-optimal and slow charging cost time.

Consequently, electric car improvements in term of performance cannot always be translated for EA.

2.1.2 Aerial Vehicle

Surprisingly, electric aircrafts date back as early as the 1880s, when the Tissandier brothers first flew an electrically powered airship [14]. Though, it is only in the 1970s that EA returned when Robert Boucher was commissioned by the Defence Advanced Research Project Agency (DARPA) to develop multiple experimental aircraft [15]. This led to the development of a solar-power aircraft named Sunrise II [16]. EA later gained in popularity after the famous flights of Helios in the early 2000s [15]. Helios was a flexible flying wing power by solar panels and hydrogen fuel cells [17].

These advancements soon pushed for the creation of a series of experimental aircraft with various energy storage. For example, Boeing announced their concept SUGAR Volt, a hybrid hydrogen-powered aircraft initially brainstormed in 2006 [18].

Airbus followed with a different approach announcing their aim to reduce by half their carbon emission by 2050, and to develop a fully electric regional aircraft by 2030 [19].

To start the process, they have built and successfully tested the E-Fan 1. However, the

first successful commercial electric flight was achieved by a much smaller company,

(19)

named Harbour Air, who tested their EA in December 2019 [20]. Thus, proving that batteries may be the best solution for storing energy. Since Helios, other solar-powered aircraft has been developed and tested, but with limited practical success. Impulse 2 has completed an impressive round of the world tour fully powered by the sun [21].

But due to its size, limited payload and speed, solar-powered EA will require further development to achieve mass adoption.

Beside these experimental aircraft, a new wave of EA is taking shape with the development of Urban Air Mobility (UAM). The concept is simple, make cities transport airborne, with a network of air taxi to reduce congestion and pollution, and to follow the conversion from fuel to electric. Many companies are in the race to make UAM a reality with some concepts in advanced development stages. These concepts are mostly VTOL or multicopter configurations for urban operations, and fixed-wing for regional flights. Although these designs all have their unique shapes, they all share a key feature. They are all powered with batteries.

Table 2-1: Recent EA designs, [22] [23] [24] [25] [26] [27] [28] [29]

Boeing Hyundai

Lilium Vahana

Ehang 184 Volocopter

Harbour Air Alice

(20)

Numerous studies [30] [31] [32] have already been conducted to research optimal means of powering EA. The common key elements of these studies are the energy and power requirement of EA, often pointing out that current battery technology is not yet adequate to meet these requirements [30]. The main issue arises from the short take-off phase, which requires significant power in comparison to cruise. Therefore, a single battery configuration aircraft design makes a compromise between battery optimisation to account for the take-off and cruise. This has led to the proposed hybrid battery system to both reach energy and power requirements with two separate batteries instead of under optimised a singular battery.

Figure 2-1: Hybrid battery powertrain [30] [31]

Although these studies [30] [31] [32] look intensively into the conceptual design of EA, ranging from full-electric, hybrid-electric and hybrid, they do not offer specific insight into battery’s weight reduction and means of realizing a hybrid-electric system.

To cover this gap, the presented research will look specifically into hybrid-electric system’s performance in terms of battery weight reduction and propose a circuit design.

2.2 ENERGY STORAGE AND BATTERIES

The birth of battery technology can be traced back to as early as 1800 when Volta

invented the voltaic pile [33]. The battery was then made of a pile of copper and zinc

discs, separated by a layer of cloth soaked in brine. Since then, batteries have been

drastically improved, using a variety of materials and chemicals to satisfy the general

industries demand. However, current batteries are still below their theoretical potential

[34], and other energy storage devices like fuel-cells or supercapacitors have their

(21)

limitation. Further advancements have yet to be before the aviation’s industry demand is satisfied.

2.2.1 Technology Review

Energy storage comes in all forms and shapes, storing energy either in solid, liquid or gas solutions. When it comes to comparing various energy storage, the two main factor remains the same. Specific energy, which is the energy to weight ratio, and the specific power, which is the power to weight ratio. In simple aeronautical term, the specific energy links to the flight endurance of the vehicle, and the specific power links to the acceleration of the vehicle.

Some of the most used batteries across industries are lithium-ion (Li-Ion) and lithium-ion polymer (Li-Po) [35]. Li-Ion uses a liquid electrolyte solution, while Li- Po uses solid electrolyte. Li-Ions are widely used as batteries for electric cars, power tools and powerbanks since they are cheaper to manufacture, offer a higher energy density and longer lifetime. On the other hand, Li-Pos are widely used to power UAVs and cell phones for examples as they have a much higher specific power and can be more easily shaped.

Another common mean of energy storage is done with supercapacitors. A very high Farad rates capacitor allows short-term high-power charge and discharge, making them ideal for short-term energy storage or burst power delivery. For example, they are commonly used in automobiles and buses for regenerative braking [11]. However, even the latest supercapacitor advancements barely reach 10Wh/kg [36] compared to 200Wh/kg of a regular Li-Ion battery, thus greatly limiting their usability for energy- intensive applications.

On the other end, hydrogen fuel-cells are known to have high specific energy, with up to 10 times greater than Li-Ion (Figure 2-2: Ragone chart for various energy storage types . Hydrogen gas being the lightest element of all makes it attractive as a mean to store energy for weight-restricted applications such as EA. But unfortunately, storing hydrogen has been proved to be challenging, and the power performance of fuel-cells are much lower than Li-Ion batteries.

Other possibilities include methods of collecting power in flight, such as using

solar panels, tethered to ground station or even by beaming power with a laser, but

these solutions are impractical. Rather, the solution to power full-electric aircraft relies

(22)

on finding a compromise between specific energy and specific power or to use multiple storages.

2.2.2 Upcoming advancement

One of the most anticipated upcoming battery is the lithium-air (Li-air), which has a theoretical energy storage of 11kW/kg [34], compared to 200W/kg for a Li-ion.

Such battery could outperform gasoline in specific energy, but the technology is far from being ready.

On a shorter-term, another awaited development in battery technology is the all- solid-state battery (ASSB). They are projected to hold as much as twice the capacity of a regular Li-ion cell, but they are expected to be more costly and have limited power performance due to the higher internal resistance. Nevertheless, Toyota has announced its production and demonstration at the 2020 Olympics in Tokyo, proving that the technology seems to be soon available [37]. If solid-state keeps its promise of doubling energy density of batteries, EV could greatly beneficiate from it.

2.2.3 Limitation

Figure 2-2: Ragone chart for various energy storage types [37]

As mentioned earlier, the selection of a battery system for any EV inevitably

includes a trade-off between high specific energy and high specific power, which can

be visualized with the Ragone chart above.

(23)

2.3 LOAD SWITCHING CIRCUIT

If two batteries with different voltages are placed in parallel, the battery with the lower voltage will receive current from the higher voltage, which can lead to the battery malfunctioning, overheating or even exploding. Therefore, switching between two power sources, or batteries, requires a circuit to prevent the batteries of affecting one an another.

2.3.1 Mechanical

The most straightforward way of realising the circuit would be with a manual, mechanical switch where an operator could select between two different battery sources, but this would be impracticable. One step further could lead to the use of relays, where a mechanical switch is activated digitally. However, relays do not scale well as they are relatively heavy, costly and can carry a limited amount of current. This would not be an ideal solution for large scale EA application.

2.3.2 Digital

Looking at a digital approach, there are two choices currently used in a wide range of applications to serve as a load switch. First are the thyristors, but they are not easily scalable, and they are costly. The second choice, which is the most used, are transistors, such as BJTs, JFET, MOSFETs and IGBTs. When it comes to power applications, however, MOSFET and IGBTS are the usual solutions as they are relatively cheap, scalable and can share large current [39]. Though, N-Channel MOSFET traditionally offer the lowest resistance of all option, therefore meaning it is the most efficient solution for high power application.

Figure2-3: Switch on the left, Relay on the right (symbols taken from Sparkfun)

Figure 2-4: From left to right: Thyristors, BJT, JFET, MOSFET, IGBT symbol

(symbols taken from Sparkfun and EasyEDA)

(24)

2.3.3 Solving the body diode problem

A MOSFET comes with an internal body diode, which in load switching applying causes an issue. Because of the reverse body diode, current can flow back to the power source, which is as previously discussed, undesirable. The simple solution to prevent the body diode problem is to use a dual MOSFET configuration. With this approach, the counter facing diodes prevent any undesired current flow, but to the cost of doubling the power loss, component count and cost.

Figure 2-5: Dual MOSFETs configuration to prevent reverse current [40]

2.4 SUMMARY AND IMPLICATIONS

On December 10, 2019, the first fully electric commercial airplane completed a successful test flight, marking the beginning of a new electric aviation era [20]. Taken in consideration with other models such as Alice, EHANG 184 and Hyundai’s Smart Mobility Solution, the UAM movement seems to be right around the corner. However, battery technology is not yet ready to see electric air taxis take over the sky and let alone large-scale commercial airplanes. Further improvements in energy storage must be achieved before the electric aircraft revolution can become a practical reality.

Besides waiting for battery technology to improve, previous studies [30] [31]

[32] has already pointed out to means of improving current performance by utilizing and hybrid battery systems. Such a proposed system would be composed of a small, power-dense module for take-off phase, and a larger energy-dense module for the cruise phase. However, these studies have not investigated this solution in depth. For this reason, the proposed thesis will conduct a more direct approach to validate this solution based on measurable results.

The final parametric analysis will examine how various variables influence the

answer to the research question and determine to which extent such solution can be

beneficial, or not. Through the methodology of the conceptual design of a hybrid

battery module, this thesis can contribute to the development of future electric aircraft.

(25)

Chapter 3: Design and Simulation

To best answer the research question, a step by step approach is undertaken to start from a 3D model of an aircraft and to end up with a conceptualized battery module. This approach can then be automated in a loop to gather data points for later analysis to corroborate to the research question.

This step by step approach will primarily consist of two tools. First, OpenVSP is used to evaluate the aerodynamic aspects of the aircraft, and secondly, a MATLAB application will be created to use the previously found aspects and turn it into a conceptualized battery module.

This chapter first discusses the methodology and design of the research.

Secondly, the tools used are described and justified. Then, the procedure and timeline of the research are presented. Section 3.4 discusses the parametric analysis approach, and section 3.5 discusses the MATLLAB application in more detail as well as its simplifications and limitations.

3.1 METHODOLOGY AND RESEARCH DESIGN 3.1.1 Methodology

There exist a broad range of different aircraft configurations and user cases, thus making the conceptual design of a single powertrain impractical. Instead, this research will conduct three case studies of three main categories and sizes of aircraft to best cover the broad range of powertrain design requirements. Each case study will investigate if a hybrid battery system is tangible for the specific configuration.

The first evaluated configuration will be a fixed-wing aircraft. They conventionally are the most efficient for long-range and at carrying heavy payloads.

The case study for this configuration will be on the design of Alice from Eviation, a

promising regional passenger airplane.

(26)

14 Chapter 3: Design and Simulation

The second evaluated configuration will be a multirotor aircraft. Their modern design, derived from consumer UAVs (drones), gives them the agility required to navigate in dense urban areas. This configuration has been often used to represent the air taxi trend. The case study for this configuration will be on the design of EHANG184 from Ehang, a single-seater autonomous air taxi.

Figure 3-2: EHANG 184 [26]

The third and last evaluated configuration will be a VTOL aircraft. They combine the efficiency of fixed-wing with the agility of multicopter. The case study for this configuration will be on the design of a humanitarian rescue UAV from A3T, the student drone team from the University of Twente.

Figure 3-1: Alice aircraft from Eviation [41]

(27)

Figure 3-3: VAU, designed by members from A3T (courtesy of A3T)

The step by step approach is relatively straightforward. The goal is to find the drag coefficient of the aircraft relative to different phases of flight, input essential parameters in the MATLAB application and it outputs results. These steps are:

1. Model the aircraft in OpenVSP.

2. Use the parasite drag tool to find 𝐶

_𝐷0

at for each flight phase.

3. Use VSPAERO to find the lift coefficient at each segment.

4. Calculate the average to find the total lift coefficient.

5. Calculate the Oswald factor.

6. Calculate the drag coefficient.

7. Repeat steps 1 to 7 for each flight phase.

8. Input the drag coefficient and the required parameters in the application.

With this approach, the application will be able to determine how much power

and energy is required to complete the mission profile of the aircraft and conceptualise

a battery module. Using the application, the process can be automated to obtain various

data point from variables and use it to complete the analysis.

(28)

3.1.2 Research Design

There are countless variables that come into play when studying aircraft performances. However, this study will focus solely on the principal factors influencing power and energy consumption to avoid going out of scope. The first three variables are relative to the design phase of the aircraft, and the five remaining are relative to the mission profile. The list of variables used in this study is presented below:

Table 3-1: Variables studied

Variable Classification Expected impact

Drag

Coefficients

Dependent on aircraft’s design Important impact as it is directly linked to the force drag formula Surface area Dependent on aircraft’s design Important impact as it is directly

linked to the force drag formula Motor

Efficiency

Dependent on the motor’s design, cruise speed & altitude

Important impact as it is converting power to force

Speed Independent Most important impact since it is exponential to the force drag Range Independent Major impact as it relates to

energy requirements

Altitude Independent Minor impact as air density slightly varies with altitude Acceleration Independent Major impact as it relates to

power requirements

TOW Independent Major impact as it relates both to

energy and power requirements

3.2 SOFTWARE AND INSTRUMENTS

The study will principally make use of two software. OpenVSP (Appendix A) is used to cover the aeronautics, and MATLAB is used for the simulations. Besides, two other software will be employed, Excel and Tinker CAD (Appendix B).

First, the study is founded on OpenVSP, a parametric aircraft geometric tool

designed by NASA engineers [41]. This tool allows the virtual creation testing of any

aircraft configuration. Once the model is built, the first two variables mentioned in

section 3.1.2 can be derived using OpenVSP.

(29)

The core of the research however will be completed using MATLAB, as it offers rapid and reliable simulations. An application interface will be developed to facilitate the user interface and thus making this tool readily available for further studies by other parties.

Beside this two software, Excel can be used as a tool to sum outputs results from OpenVSP to parameters for the application. Concerning the circuit, rudimentary simulations are run on TinkerCAD for validation.

3.3 PROCEDURE AND TIMELINE

The overall research was structured on a standard project management workflow, conducted in 5 phases. The Gant chart can be found in the Appendix C.

• Phase 1: Research Conception and Initiation marks the beginning of the study. This phase is dedicated to research and review the literature, form the hypothesis and initialize the overall planning.

• Phase 2: Research Definition and Planning serves to define the thesis statement, to outline the research scope and goals and to decide on the experimental methodology. The last step of this phase is to finalize the planning of the remaining three phases.

• Phase 3: Modelling and Testing form the core of the study. Throughout this phase, the time is first allocated to realize the MATLAB application, to create the 3D models, to run the simulation and to realize a load switch circuit prototype. This phase ends by testing and collecting data from the application and the circuit.

• Phase 4: Data Analysis investigates the results obtained from the previous phase. The analysis processes the data in view to uncover meaningful information to confirm or reject the thesis statement. For this thesis, the analysis should be able to determine if a weight reduction can be obtained, and by how much.

• Phase 5: Research closure finalizes the study by summarizing in a report

the integrity of the research, the key findings and points out to future further

study opportunity. The final presentation marks the completion of the thesis.

(30)

3.4 ANALYSIS

The selected method for analysing the data is to complete a parametric analysis for each study cases. A parametric analysis allows to isolate and evaluate the impact of each variable, thus providing an understanding toward its direct influence on the resulting battery weight. In more details, each variable from Table 3-1: Variables studied will be varied from -25% to +25% at 5% intervals and plotted in a radar chart to present the resulting battery weight variation impact.

Since the objective of the research is to ultimately reduce the weight of an EA’s battery, the first step is to understand the causes behind the battery’s weight and requirements. Using a parametric analysis paired with study cases should, theoretically, pinpoint the causes delimiting the battery’s performance.

Since each configuration has different performance and requirements, it is expected that a hybrid battery system will be more useful in some cases than others.

For example, since Alice has a large battery to achieve its long-range, the battery should be able to cope with the power requirement for take-off, thus not benefiting from a hybrid solution. For Ehang, it is expected to have a minimal weight reduction since the EA uses direct power to sustain flight instead of wings. Therefore, the power requirements will be high throughout the flight. However, the hybrid battery system is expected to have the most impact for VTOL configuration, since take-off requires a high-power burst and cruise requires minimal power.

Model Configuration Payload Typical Range

Expectation

Alice Fixed wing 9 passengers and 2 crew members

1000km No benefit Ehang

184 Multicopter 1 passenger or 100kg

40km 10% to 20% weight reduction

VAU VTOL 2.5kg 30km 20% to 30% weight

reduction

3.5 THE APPLICATION

Instead of manually calculating each study case, a MATLAB application is

written to serve as a tool to conduct the simulations (Appendix E). Through the

application and its simulations, it will be possible to evaluate how much energy and

(31)

power is required for different aircraft configurations and mission profiles. Then, the application will be able to propose a hybrid battery module configuration from the calculated energy and power requirements. The application also possesses a sweep function that evaluates the impact of each selected variables separately through a parametric analysis. This sweep function can then highlight the main factors responsible for validating the research question.

3.5.1 Front-end

The front-end of the application is a simple window to input all variables required to fulfil the simulations and visualize the results. The application is split over two tabs, the first being the “Energy & Power Profile” and second for the “Battery Profile”. Each tab holds multiple subsections to classify the data. The subsection in blue are for input, and the subsection in grey are for outputting the results.

3.5.2 Back-end

Once all parameters are entered and the “calculate” button is clicked, the program will execute a series of calculations to deduct the energy and power requirements of the aircraft.

Figure 3-4: Energy & Power Profile on the left and Battery Profile on the right

(32)

3.5.2.1 General structure

The program simulates a hypothetical flight by calculating data points at a one second intervals through each flight phase. The data points are recorded in data structure under each of their respective variable. This data structure consists of the following variables:

• Time, to mark the timestamp in second

• Altitude, in meters

• Range, to mark the current distance travelled in meters

• Speed, in meters per seconds

• ForceDrag, to specify the force induced by drag in Newtons

• ForceTotal, to specify the thrust required by the propulsion in Newtons

• Power, to specify the power required by the motors in Watts

• Efficiency of the motors in g/W

• Phase, to indicate which is the current phase.

The first phase is the take-off. During that phase, the program loops and calculate each variable until the final take-off speed is achieved. For fixed-wing, the final take- off speed is 1.1 the stall speed [42], and for both multicopter and VTOL, it is until climb rate speed is achieved. Since the time step is every second, the speed is calculated by adding the acceleration at each step. For the VTOL, a transition loop is added to accelerate the aircraft up to 1.1 the stall speed.

The second phase is the climb. During that phase, the program loops and calculate each variable until the vehicle reaches cruise altitude and speed. The altitude is calculated by adding the climb rate at every step. Furthermore, a constant acceleration is calculated for the vehicle to go from 1.1 stall speed until cruise speed for the fixed-wing and VTOL, and to go from 0km/h (horizontal speed) to cruise speed for the multicopter.

The third phase is the cruise. Before starting the calculation loop, the program

first checks the distance the vehicle will travel in it descends. The cruise distance is

then equal to the input range, minus the distance covered during climb and descend.

(33)

Once the cruise distance is known, the program loops and calculate each variable until the vehicle achieve the cruise distance.

The final phase of the flight is the descend and landing. The descend follows the same steps as the climb phase but using the descend rate and final speed of 1.3 stall speed [42] for the fixed-wing and VTOL. The landing phase is omitted in this study as the vehicle usually uses parasite drag or relatively minimal energy to stop the vehicle, thus it is not expected to influence the results.

During each loop, the program utilizes three functions where the variables from the current status of the aircraft are inputted to calculate the following status.

3.5.2.2 Functions

The first function is the barometric calculator. This function takes as an input the altitude and outputs the air density. The barometric calculator uses three different standard formulas [43] depending if the aircraft is either in the troposphere, lower stratosphere or upper stratosphere.

When in the troposphere, between 0 to 11km altitude, the formulas for the temperature (T) and pressure (P) are:

𝑇 = 15.04 − 0.00649 ∙ ℎ

𝑃 = 101.29 ∙ [ 𝑇 + 273.1 288.08 ]

5.256

When in the lower stratosphere, between 11km and 25km, the formulas for the temperature and pressure are:

𝑇 = −56.46

𝑃 = 22.65 ∙ 𝑒

(1.73−0.000157∙ℎ)

When in the upper stratosphere, above 25km, the formulas for the temperature and pressure are:

𝑇 = −131.21 + 0.00299 ∙ ℎ

𝑃 = 2.488 ∙ [ 𝑇 + 273.1 216.5 ]

−11.388

The air density can then be calculated using the temperature and pressure:

(34)

𝜌 = 𝑃

0.2889 ∙ (𝑇 − 273.1)

The second function is the drag calculator. At every step, the air drag is calculated using as input the drag coefficient, the altitude, the speed and the surface area of the vehicle respective to its current status. The air pressure is calculated using the previous function and provides the air density for the drag formula. This function then returns the force caused by drag [44]:

𝐹

_𝐷

=

¹

2

𝜌𝑣

²

𝐶

_𝐷

𝐴 Equation 3-1 The last function is the power calculator, which calculate the power required to propeller the vehicle. The function takes as input arguments the total force required and motor parameters. The total force is calculated by adding the force caused by drag with the force required for the acceleration or deceleration of the vehicle.

Then, the program uses the motor parameters of the efficiency versus thrust at both 50% and 100% power to establish a linear function. Using this function, the power can be estimated for any given thrust requirement. However, the program implements two corrections factors before translating to the power requirement. First, the thrust of the motors is corrected to account for the lower air density in altitude. This is done by multiplying the thrust parameter with the ratio of the air density at ground level to the air density at the current altitude. Secondly, the force required is corrected for the loss of force due to the vehicle’s velocity. As seen below, the given force of a motor and its propeller can be model with this formula, with 𝜌 being the air density, 𝐴 the area covered by the propeller, 𝑣

_𝑒

the exit air velocity and 𝑣

₀

the intake air velocity [45].

𝐹 =

¹

2

𝜌𝐴(𝑣

_𝑒²

− 𝑣

₀²

) Equation 3-2 When testing is done or when the vehicle is static, the formula can be simplified as below:

𝐹

_𝑠

= 1 2 𝜌𝐴𝑣

_𝑒²

Thus, the actual force provided by the propeller when the aircraft is at 𝑣

₀

is then:

𝐹

_𝐷

= 𝐹

_𝑠

− 1

2 𝜌𝐴𝑉

₀²

(35)

The final power can then be calculated provided with the correction of the air density at altitude and the extra force required for each motor due to the speed of the vehicle.

3.5.3 Battery Weight Calculation

Once the program as calculated all the data points, the power requirement of each point is integrated over time to provide the energy consumption for all phases of flight. As for the power, the highest power requirement of each phase is set as the phase’s power requirement.

The program then loops through the battery list to calculate the optimal main battery weight to cover cruise and descend, as well as to calculate the optimal auxiliary battery to cover take-off and climb requirements. For each battery in the list, the program checks the estimated weight of the battery module based on the energy density and power density of the listed battery, then select the highest resulted weight and store this value in an array. After cycling through the list, the program outputs the lowest weight result of the array as the optimal battery configuration. The process is subsequently repeated but for a singular battery solution to serve as reference weight.

3.5.4 Simplifications and Limitations

As previously mentioned, it was required to make multiple simplifications due to the complexity of aerodynamics and the time constraint. Theses simplifications largely originate from the flight profile for each study case and the efficiency of the motors.

For this simplified study, the flight profile of the fixed-wing model starts with the take-off phase, where the aircraft accelerates until it reaches 1.1 stall speed. The vehicle then climbs and accelerates until it reaches cruise speed and altitude. After the cruise, the aircraft descends until reaches ground and 1.3 stall speed. However, real-

Figure 3-5: Simplified Flight Profile of a Fixed Wing

(36)

life profiles are more complex and include for example more detail flight phases such as departure, approach and taxi, but this is beyond the scope of the project.

The multicopter and VTOL have slightly different flight profile. For the multicopter and VTOL, the take-off and landing are vertical, for which the change in altitude is considered. For the VTOL, a transition stage is added to change between multicopter mode to fixed wing.

The other main simplification is the for the motor efficiency within the power calculator function. The thrust reduction due to the lower air density and dynamic state of the aircraft is complex and would require advance fluid computation and analysis to be done accurately. But for the scope of the project, a direct ratio to adjust for the air density and a force deduction corrected with a factor of 0.25 is estimated to be sufficiently accurate. However, these two simplifications do reduce the accuracy of the simulations and would need to be further improved for later studies.

Besides these two main simplifications, other factors such as the weather conditions and wind current are not considered. As for all the parameters, approximations were taken when the exact number could not be found or calculated.

Moreover, the list of battery used for the simulation is limited to Kokam Lithium Ion cell [46] as their inventory adequately covers battery ranging from ultra-high energy density to ultra-high power density.

Figure: 3-6: Simplified Flight Profile of a Multicopter on the left and VTOL on the right

(37)

Chapter 4: Results & Validation

This chapter presents the results from the parametric analysis and the calculated battery weight for each study cases. In every simulation, the parametrically analysed variable is varied from -25% to +25% with increments of 5% to graph its impact on the resulting battery weight between a hybrid and singular solution. The reference values for these simulations are presented in the Appendix D.

4.1 ALICE

Table 4-1: Weight and capacity comparison

Weight (kg) Capacity (kWh)

Singular 147.96 26.51

Singular - Weight Adjusted 158.78 28.42

Hybrid 158.78 35.59

1000km for 3.7T of Battery

1056km for 3.7T of Battery 1000km for 3.2T of

Battery

0,00 200,00 400,00 600,00 800,00 1000,00 1200,00

Range (km)

Range Comparaison between Hybrid and Singular Battery Solution for Alice

Singular Solution Singular Solution Ajdusted to weight of Hybrid Hybrid Solution

Figure 4-1: Range performance relative to battery weight

-5%

0%

5%

10%

Drag Coefficients

Surface Aera

Motor Efficiency

Speed (Adjusted by factor of 0.1 to fit chart) Range

Altitude Acceleration Take Off Weight

Hybrid and Singular Battery Weight Variation Relative to 5%

Variation per Variable for Alice

Hybrid Singular

Figure 4-2: Results of the parametric analysis for the fix wing configuration

(38)

26 Chapter 4: Results & Validation

4.2 EHANG 184

Table 4-2: Weight and capacity comparison

Weight Capacity

Singular 147.96 26.51

Singular - Weight Adjusted 158.78 28.42

Hybrid 158.78 35.59

-5%

0%

5%

10%

Drag Coefficients

Surface Aera

Motor Efficiency

Speed Range

Altitude Acceleration Take Off Weight

Hybrid and Singular Battery Weight Variation Relative to 5%

Variation per Variable For Ehang184

Hybrid

Singular

49km for 159kg of Battery 34km for 158kg of

Battery 30km for 148kg of

Battery

0,00 10,00 20,00 30,00 40,00 50,00 60,00

Range (km)

Range Comparaison between Hybrid and Singular Battery Solution

Singular Solution Singular Solution Ajdusted to the battery weight of Hybrid Hybrid Solution

Figure 4-4: Results of the parametric analysis for the multicopter configuration

Figure 4-3: Range performance relative to battery weight for the multicopter

(39)

22,5 24 25,5 27 28,5 30 31,5 33 34,5 36 37,5 Energy Density Difference -6% -5% -5% -4% -4% 25% 25% 25% 25% 25% 25%

Weight Difference 13% 13% 12% 10% 10% 7% 3% -1% -5% -9% -12%

-15%

-10%

-5%

0%

5%

10%

15%

20%

25%

30%

35%

Difference (%)

Range (km)

Weight And Energy Density Difference of An Hybrid Solution Compared With a Singular Battery For Various Range of Ehang184

75 80 85 90 95 100 105 110 115 120 125

Energy Density Difference 24% 25% 25% 25% 25% 25% -4% -4% -5% -5% -5%

Weight Difference -16% -12% -7% -2% 3% 7% 10% 10% 12% 12% 12%

-20%

-10%

0%

10%

20%

30%

40%

Difference (%)

Speed (km/h)

Weight And Energy Density Difference of An Hybrid Solution Compared With a Singular Battery For Various Cruise Speed of

Ehang184

Figure 4-5: Battery weight reduction achieved at different cruise speed for Ehang184

Figure 4-6: Battery weight reduction achieved at different range speed for Ehang184

(40)

28 Chapter 4: Results & Validation

4.3 VAU

Table 4-2: Weight and capacity comparison

Weight (kg) Capacity (Wh)

Singular 1.37 204.06

Singular – Weight Adjusted 1.59 226.35

Hybrid 1.59 312.14

-5%

0%

5%

Drag Coefficients

Surface Aera

Motor Efficiency

Speed Range

Altitude Acceleration

Take Off Weight

Hybrid and Singular Battery Weight Comparaison Relative to 5%

Variation per Variable for VAU

Hybrid

Singular

73.64km for 1.59kg of Battery 47km for 1.59kg of

Battery 40km for 1.37kg of

Battery

0,00 10,00 20,00 30,00 40,00 50,00 60,00 70,00 80,00

Range (km)

Range Comparaison between Hybrid and Singular Battery Solution

Singular Solution Singular Solution Ajdusted to the battery weight of Hybrid Hybrid Solution

Figure 4-7: Results of the parametric analysis for the VTOL configuration

Figure 4-8: Range performance relative to battery weight for the VTOL configuration

(41)

Chapter 5: Parametric Analysis

A significant amount of data was collected through the parametric analysis of the 8 variables tested for each study case. Since these results originate from simulations based on multiple simplifications and assumptions, the analysis will not focus on the numerical significance of the results. Instead, the analysis will focus on the relations between these results and develop a conclusion from thereon.

This chapter is divided into three sections. The first section will review the results and their underlying findings, and the second section will evaluate the resulting aircraft performance. The last section will review a proposed load switch circuit approach.

5.1 THE RESULTS 5.1.1 Alice

As it was expected, a hybrid battery solution for the long-range fixed-wing proved not to be beneficial and even increased the overall weight of the battery. This can be explained due to the already large battery required to keep the aircraft flying for at least 2 hours. Since the battery has a relatively large capacity, it can cope with the power-intensive phase at the take-off. Dividing the battery into a hybrid solution would only reduce the overall energy density, thus making the battery heavier.

From the Table 3-1: Variables studied, variations of each variable gave similar battery weight variation for both the hybrid and singular solution, except for the take- off weight and acceleration for which the hybrid is more susceptible. This difference can be attributed to the hybrid solution requiring a singular battery to provide the acceleration power for the take-off phase, thus optimising this battery for power density instead of energy density. Moreover, the drag coefficient, surface area and range variations did closely match with the variation of the battery weight, thus outlining that these variables have a direct impact with the battery weight.

Surprisingly, however, variation of the speed did significantly influence the

weight of the battery. At 75% of the reference cruise speed, the battery weight for both

the hybrid and singular solution was cut in half. On the other hand, the battery weight

went up by 5 folds when simulated with 125% the reference cruise speed. This may be

(42)

30 Chapter 5: Parametric Analysis

explained by the equations of drag (𝐹

_𝐷

=

¹

2

𝜌𝑣

²

𝐶

_𝐷

𝐴 Equation 3-1) and propeller force (𝐹 =

¹

2

𝜌𝐴(𝑣

_𝑒²

− 𝑣

₀²

)

Equation 3-2). At low speed, the drag force of an aircraft remains minimal and the force provided by the propeller is the greatest. However, these change at a higher speed. Since velocity is squared, for example, doubling the velocity of the aircraft will cause four times more drag and thus requires four times more thrust. This also adds to the propeller being negatively affected by the aircraft velocity, as the propeller performance decreases as the aircraft goes faster.

It seems as the performance of Alice could potentially be improved by reducing the cruise speed to achieve the same distance with less energy. But it is most likely that the velocity correction for the motor efficiency in the simulation is not accurate and that Alice’s design is already optimised. Besides this possible inaccuracy, the resulting battery weight of 3.2 Tones closely matches the actual Alice’s battery of 3.7 Tones. Therefore, the simulation seems to be sufficiently accurate for this study case.

5.1.2 Ehang184

Overall, the results also pointed out that a hybrid solution for Ehang184 is not always beneficial. In cruise speed of 75 to 90km/h and in the operational range of 33km to 48km, a hybrid solution did provide a battery weight reduction relative to a singular solution. Unfortunately, this is because the singular battery solution was heavier compared to the reference value for lower speed and longer range, and so does not give a significant reduction compared to the reference weight. This can likely be attributed to the design already being optimised for this speed and range.

Unexpectedly, however, a hybrid solution resulted in a 25% energy density increase for only 7% weight increase, thus substantially improving the aircraft’s performance for a minimal weight cost. This result points out that a hybrid solution inadvertently cost additional weight compared to a singular battery, but with the benefits of a higher energy density.

As anticipated, the parametric analysis of the multicopter gave distinct results

compared to the fixed-wing. Since multicopter does not travel as fast as fixed-wing,

variables directly linked to the drag force, as the drag coefficients and surface area,

have minimal impact on the battery weight. Another interesting result, however, is that

a hybrid solution seems to be more stable to speed and range variation. But for both

(43)

solutions and increase in take-off weight can be directly linked to an increase in the aircraft’s battery weight, thus outlining that TOW is the most impactful variable for this configuration.

Simulation wise, the estimated weight of 148kg paired with a person of 75kg would leave about 100kg to realise the construction of the vehicle. Although there is no mention on the vehicle’s battery weight online, it is safe to assume that the simulation likely matches the real value. Furthermore, the hybrid solution would give an additional 20km range on top of the original 30km range for only 10kg battery weight extra. This would be achievable aircraft optimisation.

5.1.3 VAU

Once again, the parametric analysis profile is unique to the aircraft’s configuration. For VAU, a hybrid solution seems to be generally more stable to multiple variable variations compared to a singular solution. For exception, a hybrid solution is more susceptible to variation of take-off weight and take-off acceleration.

But for both variables, a variation of 5% still results in less than 5% battery weight variation.

Surprisingly again, the hybrid solution did improve the aircraft’s performance but not as expected. For an 11% weight increase, the hybrid solution gives a 38%

energy density increase. Thus, further supporting the result that a hybrid solution can improve an aircraft’s range but to the cost of a slight weight increase.

Finally, the simulations appeared to be accurate as of the estimated 1.4kg of battery matches to the 1.2kg of the real design. Furthermore, the results conclude that an increase of 200gr in battery could improve the operational range from 40km to 73km, which would be an acceptable weight increase.

Conceptual design of hybrid battery modules for optimising electric aircraft performance